As a first step toward developing a system of outcome-based reimbursement for nursing homes, this study will explore techniques for establishing accurate and reliable prognoses for long-term patients. The study will utilize a series of repeated observations by a multidisciplinary team on approximately 500 nursing-home patients receiving good care. The team will gather data, develop prognoses on each patient, and compare these to actual patient outcomes. The prognostic process will be modeled mathematically using this same data base. Because the function of nursing-home patients is a complex mixture of physiological, psychological, and social factors, a portion of this project will be devoted to reducing this multidimensional matrix to a unidimensional scale. An additional source of prognostic judgments will be gathered from a panel of national experts who will review case histories gathered by the team to estimate expected patient outcomes. Each component of this study is thus separate but very much dependent on the other portions.